b'Accelerating and improvingMachine learning-driven uncertainty quantification methods accelerate the reliability of low failurehigh-fidelity model failure evaluation of nuclear reactor technologies.probability computationsS afety evaluation of advanced nuclear technologies is a critical step toward ensuring the operational reliability and the sustainability of nuclear to support the efficientenergy. However, computationally expensive finite element models safety evaluation andwith many uncertain input parameters are often employed to represent nuclear technologies and simulate failure. The resulting failure probabilities tend to be low, deployment of advancedon the order of 10-310-7. This project researched artificial intelligence-driven reactor technologies Monte Carlo algorithms to efficiently estimate low failure probability values using computationally expensive finite element models. These artificial intelligence-enabled algorithms reduced the number of finite element model evaluations by several orders of magnitude compared to state-of-the-art methods. Three case studies were considered for evaluating the performance of the artificial intelligence-enabled algorithms: failure of the tristructural isotropic particle fuel, stresses, and thermal failure modes in a heat pipe microreactor, and embrittlement PROJECT NUMBER:failure of a reactor pressure vessel. Advanced Monte Carlo algorithms were also 21A1050-114FP implemented in the MOOSE so that the broader nuclear community can make TOTAL APPROVED AMOUNT:use of these efficiency improvements. Our research enabled the consideration of $999,000 over 3 years high-fidelity computational models for the failure analysis of the three case studies, whereas the state-of-the-art relies on low-fidelity models. We published these PRINCIPAL INVESTIGATOR:new developments in scientific journals and presented the new developments at Som Dhulipala international conferences and invited lectures and talks. This project supported CO-INVESTIGATORS: the efficient safety evaluation of advanced reactor technologies, leading to their Andrew Slaughter, INL accelerated innovation, design, and optimization.Benjamin Spencer, INLChandrakanth Bolisetti, INLPeter German, INLVincent Labour, INLYifeng Che, INLZachary Prince, INLMichael Shields, Johns Hopkins University COLLABORATOR:RWTH Aachen University, Germany 42'